Journal of Emerging Trends in Engineering and Applied Sciences: Research Areas

Article Title: A Novel Framework for Big Data Processing in a Data-Driven Society
by Otuonye Anthony I., Onyechere Patricia O., Njoku Obilor A.

Abstract:
Big data has continued to attract huge attention from the academia, as well as the industry and government agencies. The attraction has been due to a quick rise in the generation of large-scale and varied data, resulting from emerging increase in the use of sensor networks, social networks, and some semantic web applications. Spontaneous increase in large-scale data poses a critical challenge for today�s Data Scientists and Social Managers. This paper introduces a big data processing framework from the data mining perspective that will fully harness the potential benefits of the big data revolution and enhance effective management and processing of large-scale data. We have proposed a three-tier data mining structure for big data storage, processing and analysis from a single platform and provide social sensing feedback for a better understanding of our society. Big Data concern large-volume, complex, and growing data sets with multiple, autonomous sources. Our model is both data-driven and demand-driven from various information sources, mining and analysis. The study became necessary due to the growing need to assist governments and business agencies to take advantage of the big data technology for the desired turn-around in their socio-economic activities. We have adopted the HACE theorem in the model design which characterizes the unique features of the big data revolution: large, heterogeneous, complex, and evolving. We also adopted the Hadoop�s Map Reduce optimization strategies based on the MapReduce parallel processing framework for big data mining. There is need to revise most of our traditional data mining techniques and deploy the suggested distributed versions of big date models available to ensure profitable data analysis in our contemporary data-driven society.

Keywords: Big Data, HACE Theorem, Data Mining, Social Network, Hadoop

Download full paper